You are on page 1of 32

Political Science and the

Three New Institutionalisms

Peter A. Hall and Rosemary C. R. Taylor

96 / 6

This paper was presented by Professor Hall, a member of the


MPIFG Scientific Advisory Board, as a public lecture during the
Board’s meeting on May 9, 1996.

To contact the authors, write to Professor Peter A. Hall, Political Science Department,
Harvard University, 27 Kirkland Street, Cambridge, MA 02138, USA.

Max-Planck-Institut für Gesellschaftsforschung


Lothringer Str. 78
D-50677 Köln
Germany

Telephone 0221 / 33605-0


Fax 0221 / 3 36 05 -55 MPIFG Discussion Paper 96 /6
E-Mail mpi@mpi-fg-koeln.mpg.de ISSN 0944–2073
Home Page http: / / www.mpi-fg-koeln.mpg.de June 1996
Hall, Taylor: The Three New Institutionalisms 3

Contents

1 Introduction 5

2 Historical Institutionalism 5

3 Rational Choice Institutionalism 10

4 Sociological Institutionalism 13

5 Comparing Institutionalisms 17

6 Conclusion 22

7 Notes 25
Hall, Taylor: The Three New Institutionalisms 5

1 Introduction

The ‘new institutionalism’ is a term that now appears with growing frequency in
political science. However, there is considerable confusion about just what the
‘new institutionalism’ is, how it differs from other approaches, and what sort of
promise or problems it displays. The object of this essay is to provide some pre-
liminary answers to these questions by reviewing recent work in a burgeoning
literature.

Some of the ambiguities surrounding the new institutionalism can be dispelled if


we recognize that it does not constitute a unified body of thought. Instead, at
least three different analytical approaches, each of which calls itself a ‘new insti-
tutionalism,’ have appeared over the past fifteen years. We label these three
schools of thought: historical institutionalism, rational choice institutionalism,
and sociological institutionalism.1 All of these approaches developed in reaction
to the behavioral perspectives that were influential during the 1960s and 1970s
and all seek to elucidate the role that institutions play in the determination of so-
cial and political outcomes. However, they paint quite different pictures of the
political world.

In the sections that follow, we provide a brief account of the genesis of each
school and characterize what is distinctive about its approach to social and politi-
cal problems. We then compare their analytical strengths and weaknesses, focus-
ing on the stance that each adopts toward two issues fundamental to any institu-
tional analysis, namely, how to construe the relationship between institutions
and behavior and how to explain the process whereby institutions originate or
change.

Given the similarity in their interests, it is paradoxical that these three schools of
thought developed quite independently of each other, at least judging from the
paucity of cross-references in the literature. Until recently, there has been little in-
terchange among them. Accordingly, we ask what each might learn from the
others and, in the conclusion, what potential there is for integrating their insights.

2 Historical Institutionalism

Historical institutionalism developed in response to the group theories of politics


and structural-functionalism prominent in political science during the 1960s and
6 MPIFG Discussion Paper 96/6

1970s.2 It borrowed from both approaches but sought to go beyond them. From
group theory, historical institutionalists accepted the contention that conflict
among rival groups for scarce resources lies at the heart of politics, but they
sought better explanations for the distinctiveness of national political outcomes
and for the inequalities that mark these outcomes.3 They found such explanations
in the way the institutional organization of the polity and economy structures
conflict so as to privilege some interests while demobilizing others. Here, they
built on an older tradition in political science that assigned importance to formal
political institutions but they developed a more expansive conception both of
which institutions matter and of how they matter.4

The historical institutionalists were also influenced by the way in which struc-
tural functionalists saw the polity as an overall system of interacting parts.5 They
accepted this contention but reacted against the tendency of many structural
functionalists to view the social, psychological or cultural traits of individuals as
the parameters driving much of the system’s operation. Instead, they saw the in-
stitutional organization of the polity or political economy as the principal factor
structuring collective behavior and generating distinctive outcomes. They em-
phasized the ‘structuralism’ implicit in the institutions of the polity rather than
the ‘functionalism’ of earlier approaches that viewed political outcomes as a re-
sponse to the needs of the system.

Structural functionalism and group conflict theories had both pluralist and neo-
Marxist variants and debate about the latter played an especially influential role
during the 1970s in the development of historical institutionalism.6 In particular,
it led many historical institutionalists to look more closely at the state, seen no
longer as a neutral broker among competing interests but as a complex of institu-
tions capable of structuring the character and outcomes of group conflict.7
Shortly thereafter, analysts in this school began to explore how other social and
political institutions, of the sort associated with labor and capital, could structure
interactions so as to generate distinctive national trajectories.8 Much of this work
consists of cross-national comparisons of public policy, typically emphasizing the
impact of national political institutions structuring relations among legislators,
organized interests, the electorate and the judiciary.9 An important subliterature
in comparative political economy extended such analyses to national labor
movements, employer organization, and financial systems.10

How do historical institutionalists define institutions? By and large, they define


them as the formal or informal procedures, routines, norms and conventions em-
bedded in the organizational structure of the polity or political economy. They
can range from the rules of a constitutional order or the standard operating pro-
cedures of a bureaucracy to the conventions governing trade union behavior or
Hall, Taylor: The Three New Institutionalisms 7

bank-firm relations. In general, historical institutionalists associate institutions


with organizations and the rules or conventions promulgated by formal organi-
zation.11

In the context of the other schools reviewed here, four features of this one are
relatively distinctive. First, historical institutionalists tend to conceptualize the
relationship between institutions and individual behavior in relatively broad
terms. Second, they emphasize the asymmetries of power associated with the
operation and development of institutions. Third, they tend to have a view of in-
stitutional development that emphasizes path dependence and unintended con-
sequences. Fourth, they are especially concerned to integrate institutional analy-
sis with the contribution that other kinds of factors, such as ideas, can make to
political outcomes. Let us elaborate briefly on each of these points.12

Central to any institutional analysis is the question: how do institutions affect the
behavior of individuals? After all, it is through the actions of individuals that in-
stitutions have an effect on political outcomes. In broad terms, new institutional-
ists provide two kinds of responses to this question, which might be termed the
‘calculus approach’ and the ‘cultural approach’ respectively. Each gives slightly
different answers to three seminal questions: how do actors behave, what do in-
stitutions do, and why do institutions persist over time?

In response to the first of these questions, those who adopt a calculus approach
focus on those aspects of human behavior that are instrumental and based on
strategic calculation. They assume that individuals seek to maximize the attain-
ment of a set of goals given by a specific preference function and, in doing so, be-
have strategically, which is to say that they canvass all possible options to select
those conferring maximum benefit. In general, the actor’s goals or preferences are
given exogenously to the institutional analysis.

What do institutions do, according to the calculus approach? Institutions affect


behavior primarily by providing actors with greater or lesser degrees of certainty
about the present and future behavior of other actors. More specifically, institu-
tions provide information relevant to the behavior of others, enforcement mech-
anisms for agreements, penalities for defection, and the like. The key point is that
they affect individual action by altering the expectations an actor has about the
actions that others are likely to take in response to or simultaneously with his
own action. Strategic interaction clearly plays a key role in such analyses.

Contrast this with a ‘cultural approach’ to such issues. The latter stresses the de-
gree to which behavior is not fully strategic but bounded by an individual’s
worldview. That is to say, without denying that human behavior is rational or
purposive, it emphasizes the extent to which individuals turn to established rou-
8 MPIFG Discussion Paper 96/6

tines or familiar patterns of behavior to attain their purposes. It tends to see in-
dividuals as satisficers, rather than utility maximizers, and to emphasize the de-
gree to which the choice of a course of action depends on the interpretation of a
situation rather than on purely instrumental calculation.

What do institutions do? From this perspective, institutions provide moral or


cognitive templates for interpretation and action. The individual is seen as an en-
tity deeply imbricated in a world of institutions, composed of symbols, scripts
and routines, which provide the filters for interpretation, of both the situation
and oneself, out of which a course of action is constructed. Not only do institu-
tions provide strategically-useful information, they also affect the very identities,
self-images and preferences of the actors.13

These two approaches also supply different explanations for why the regularized
patterns of behavior that we associate with institutions display continuity over
time.14 The calculus approach suggests that institutions persist because they em-
body something like a Nash equilibrium. That is to say, individuals adhere to
these patterns of behavior because deviation will make the individual worse off
than will adherence.15 It follows that the more an institution contributes to the
resolution of collective action dilemmas or the more gains from exchange it
makes possible, the more robust it will be.16 A cultural approach, on the other
hand, explains the persistence of institutions by noting that many of the conven-
tions associated with social institutions cannot readily be the explicit objects of
individual choice. Instead, as the elemental components from which collective
action is constructed, some institutions are so ‘conventional’ or taken-for-granted
that they escape direct scrutiny and, as collective constructions, cannot readily be
transformed by the actions of any one individual. Institutions are resistant to re-
design ultimately because they structure the very choices about reform that the
individual is likely to make.17

Historical institutionalists are eclectic; they use both of these approaches to spec-
ify the relationship between institutions and action. Immergut, for instance, ex-
plains cross-national differences in health care reforms by reference to the will-
ingness of physicians’ groups to compromise with the advocates of reform, a
willingness she links, in turn, to the way in which the institutional structure of
the political system affects these groups’ expectations about the likelihood of suc-
cessfully appealing an unpalatable decision beyond the legislature.18 Hers is a
classic calculus approach. Hattam employs a similar approach, when she argues
that the entrenched power of the judiciary led the American labor movement
away from strategies that were susceptible to judicial review. However, like
many historical institutionalists, she goes farther to explore the way in which dif-
ferences in the institutional setting facing organized labor in the United States
Hall, Taylor: The Three New Institutionalisms 9

and Britain fostered trade union movements with quite different worldviews.
This kind of analysis suggests that the strategies induced by a given institutional
setting may ossify over time into worldviews, which are propagated by formal
organizations and ultimately shape even the self-images and basic preferences of
the actors involved in them.19

The second notable feature of historical institutionalism is the prominent role that
power and asymmetrical relations of power play in such analyses. All institu-
tional studies have a direct bearing on power relations. Indeed, they can usefully
be read as an effort to elucidate the ‘second’ and ‘third’ dimensions of power
identified some years ago in the community power debate.20 But historical insti-
tutionalists have been especially attentive to the way in which institutions dis-
tribute power unevenly across social groups. Rather than posit scenarios of
freely-contracting individuals, for instance, they are more likely to assume a
world in which institutions give some groups or interests disproportionate access
to the decision-making process; and, rather than emphasize the degree to which
an outcome makes everyone better off, they tend to stress how some groups lose
while others win. Steinmo, for instance, explains cross-national differences in tax
policy largely by reference to the way in which political institutions structure the
kinds of social interests most likely to be represented in the policy process.21 In
the realm of American economic policy, Weir shows how the structure of the po-
litical system militates in favor of the formation of some social coalitions and
against others.22

The historical institutionalists are also closely associated with a distinctive per-
spective on historical development. They have been strong proponents of an im-
age of social causation that is ‘path dependent’ in the sense that it rejects the tra-
ditional postulate that the same operative forces will generate the same results
everywhere in favor of the view that the effect of such forces will be mediated by
the contextual features of a given situation often inherited from the past. Of
course, the most significant of these features are said to be institutional in nature.
Institutions are seen as relatively persistent features of the historical landscape
and one of the central factors pushing historical development along a set of
‘paths.’23

Accordingly, historical institutionalists have devoted a good deal of attention to


the problem of explaining how institutions produce such paths, i.e. how they
structure a nation’s response to new challenges. Early analysts emphasized the
impact of existing ‘state capacities’ and ‘policy legacies’ on subsequent policy
choices.24 Others stress the way in which past lines of policy condition subse-
quent policy by encouraging societal forces to organize along some lines rather
than others, to adopt particular identities, or to develop interests in policies that
10 MPIFG Discussion Paper 96/6

are costly to shift.25 In this context, historical institutionalists stress the unin-
tended consequences and inefficiencies generated by existing institutions in con-
trast to images of institutions as more purposive and efficient.26

In keeping with this perspective, many historical institutionalists also divide the
flow of historical events into periods of continuity punctuated by ‘critical junc-
tures,’ i.e., moments when substantial institutional change takes place thereby
creating a ‘branching point’ from which historical development moves onto a
new path.27 The principal problem here, of course, is to explain what precipitates
such critical junctures, and, although historical institutionalists generally stress
the impact of economic crisis and military conflict, many do not have a well-
developed response to this question.28

Finally, although they draw attention to the role of institutions in political life,
historical institutionalists rarely insist that institutions are the only causal force in
politics. They typically seek to locate institutions in a causal chain that accomo-
dates a role for other factors, notably socioeconomic development and the diffu-
sion of ideas. In this respect, they posit a world that is more complex than the
world of tastes and institutions often postulated by rational choice institutional-
ists. The historical institutionalists have been especially attentive to the relation-
ship between institutions and ideas or beliefs. Goldstein, for instance, shows how
the institutional structure devised for making trade policy in the United States
tends to reinforce the impact of certain ideas about trade while undermining
others; and Weir argues that structural differences between the British and
American political systems affected the timing at which Keynesian ideas became
influential and the durability of their influence.29

3 Rational Choice Institutionalism

It is one of the curiosities of contemporary political science that a second ‘new in-
stitutionalism,’ which we term rational choice institutionalism, developed at the
same time as historical institutionalism but in relative isolation from it. Initially,
rational choice institutionalism arose from the study of American congressional
behavior. In large measure, it was inspired by the observation of a significant
paradox. If conventional rational choice postulates are correct, it should be diffi-
cult to secure stable majorities for legislation in the U.S. Congress, where the
multiple preference-orderings of legislators and multidimensional character of is-
sues should lead to rapid ‘cycling’ from one bill to another as new majorities ap-
pear to overturn any bill that is passed.30 However, Congressional outcomes ac-
Hall, Taylor: The Three New Institutionalisms 11

tually show considerable stability. In the late 1970s, rational choice analysts be-
gan to ask: how can this discrepancy be explained?

For an answer, they turned to institutions. Many began to argue that stable ma-
jorities could be found for legislation because of the way in which the rules of
procedure and committees of Congress structure the choices and information
available to its members.31 Some of these rules provide agenda control that limits
the range and sequence of the options facing Congressional votes. Others appor-
tion jurisdiction over key issues to committees structured so as to serve the elec-
toral interests of Congressmen or provide enforcement mechanisms that make
logrolling among legislators possible. In the most general terms, the institutions
of the Congress are said to lower the transaction costs of making deals so as to
allow gains from exchange among legislators that make the passage of stable
legislation possible. In short, institutions solve many of the collective action
problems that legislatures habitually confront.32

As this suggests, the rational choice institutionalists in political science drew


fruitful analytical tools from the ‘new economics of organization’ which empha-
sizes the importance of property rights, rent-seeking, and transactions costs to the
operation and development of institutions.33 Especially influential was William-
son’s argument that the development of a particular organizational form can be
explained as the result of an effort to reduce the transaction costs of undertaking
the same activity without such an institution.34 North applied similar arguments
to the history of political institutions.35 And theories of agency, which focus on
the institutional mechanisms whereby ‘principals’ can monitor and enforce
compliance on their ‘agents,’ proved useful for explaining how Congress struc-
tures relations with its committees or the regulatory agencies it superintends.36

The efflorescence of work on the American legislature that rational choice insti-
tutionalism has inspired is well represented in recent collections.37 By and large it
focuses on explaining how the rules of Congress affect the behavior of legislators
and why they arise, with an emphasis on the Congressional committee system
and the relationship between Congress and regulatory agencies. More recently,
Cox and McCubbins have attempted to shift the emphasis away from Congres-
sional committees toward the way in which political parties structure delibera-
tions. Ferejohn has begun to explore the relationship between Congress and the
courts; and a lively debate has emerged about the capacity of Congress to control
regulatory agencies.38

In recent years, rational choice institutionalists have also turned their attention to
a variety of other phenomena, including cross-national coalition behavior, the
development of political institutions, and the intensity of ethnic conflict.39 Prze-
worski, Geddes, Marks and others analyze democratic transitions in game-
12 MPIFG Discussion Paper 96/6

theoretic terms.40 Tsebelis and others explore the implications of institutional re-
form in the European Union;41 and scholars of international relations have used
the concepts of rational choice institutionalism to explain the rise or fall of inter-
national regimes, the kind of responsibilities that states delegate to international
organizations, and the shape of such organizations.42

Like all of these schools, rational choice institutionalism contains internal debates
and some variation in outlook. However, we want to emphasize four notable fea-
tures of this approach.

First, rational choice institutionalists employ a characteristic set of behavioral as-


sumptions. In general, they posit that the relevant actors have a fixed set of pref-
erences or tastes (usually conforming to more precise conditions such as the
transitivity principle), behave entirely instrumentally so as to maximize the at-
tainment of these preferences, and do so in a highly strategic manner that pre-
sumes extensive calculation.43

Second, if all schools of thought tend to promulgate a characteristic image of


politics, whether as a ‘struggle for power,’ a ‘process of social learning’ or the
like, rational choice institutionalists also purvey a distinctive image of politics.
They tend to see politics as a series of collective action dilemmas. The latter can
be defined as instances when individuals acting to maximize the attainment of
their own preferences are likely to produce an outcome that is collectively sub-
optimal (in the sense that another outcome could be found that would make at
least one of the actors better off without making any of the others worse off).
Typically, what prevents the actors from taking a collectively-superior course of
action is the absence of institutional arrangements that would guarantee com-
plementary behavior by others. Classic examples include the ‘prisoner’s dilemma’
and the ‘tragedy of the commons’ and political situations present a variety of
such problems.44

Third, one of the great contributions of rational choice institutionalism has been
to emphasize the role of strategic interaction in the determination of political out-
comes. That is to say, they postulate, first, that an actor’s behavior is likely to be
driven, not by impersonal historical forces, but by a strategic calculus and, sec-
ond, that this calculus will be deeply affected by the actor’s expectations about
how others are likely to behave as well. Institutions structure such interactions,
by affecting the range and sequence of alternatives on the choice-agenda or by
providing information and enforcement mechanisms that reduce uncertainty
about the corresponding behavior of others and allow ‘gains from exchange,’
thereby leading actors toward particular calculations and potentially better social
outcomes. We can see that rational choice theorists take a classic ‘calculus ap-
proach’ to the problem of explaining how institutions affect individual action.
Hall, Taylor: The Three New Institutionalisms 13

Finally, rational choice institutionalists have also developed a distinctive ap-


proach to the problem of explaining how institutions originate. Typically, they
begin by using deduction to arrive at a stylized specification of the functions that
an institution performs. They then explain the existence of the institution by ref-
erence to the value those functions have for the actors affected by the institution.
This formulation assumes that the actors create the institution in order to realize
this value, which is most often conceptualized, as noted above, in terms of gains
from cooperation. Thus, the process of institutional creation usually revolves
around voluntary agreement by the relevant actors; and, if the institution is sub-
ject to a process of competitive selection, it survives primarily because it provides
more benefits to the relevant actors than alternate institutional forms.45

Thus, a firm’s organizational structure is explained by reference to the way in


which it minimizes transaction, production or influence costs.46 The rules of the
American Congress are explained by reference to the gains from exchange they
provide to members of Congress. The constitutional provisions adopted by the
English in the 1680s are explained by reference to the benefits they provide to
property holders. Such examples could be multiplied. There is plenty of room for
contention within this general framework but it usually focuses on whether the
functions performed by the institution at hand are specified correctly. Thus,
Krehbiel engages the field in a lively debate about whether legislative committees
in the U.S. Congress exist primarily to provide members with gains from ex-
change or information about the outcomes of proposed legislation.47

4 Sociological Institutionalism

Independent from but contemporaneous with these developments in political sci-


ence, a new institutionalism has been developing in sociology. Like the other
schools of thought, it is rife with internal debate. However, its contributors have
developed a set of theories that are of growing interest to political scientists.

What we are calling sociological institutionalism arose primarily within the sub-
field of organization theory. The movement dates roughly to the end of the 1970s,
when some sociologists began to challenge the distinction traditionally drawn be-
tween those parts of the social world said to reflect a formal means-ends
‘rationality’ of the sort associated with modern forms of organization and bu-
reaucracy and those parts of the social world said to display a diverse set of prac-
tices associated with ‘culture.’ Since Weber, many sociologists had seen the bu-
reaucratic structures that dominate the modern landscape, in government de-
14 MPIFG Discussion Paper 96/6

partments, firms, schools, interest organizations and the like, as the product of an
intensive effort to devise ever-more efficient structures for performing the tasks
associated with modern society. The striking similarities in form taken by these
otherwise rather-diverse organizations were said to be the result of the inherent
rationality or efficiency of such forms for performing these tasks.48 Culture was
seen as something altogether different.

Against this, the new institutionalists in sociology began to argue that many of
the institutional forms and procedures used by modern organizations were not
adopted simply because they were most efficient for the tasks at hand, in line
with some transcendent ‘rationality.’ Instead, they argued that many of these
forms and procedures should be seen as culturally specific practices, akin to the
myths and ceremonies devised by many societies, and assimilated into organiza-
tions, not necessarily to enhance their formal means-ends efficiency, but as a re-
sult of the kind of processes associated with the transmission of cultural practices
more generally. Thus, they argued, even the most seemingly bureaucratic of
practices have to be explained in cultural terms.49

Given this perspective, the problematic that sociological institutionalists typically


adopt seeks explanations for why organizations take on specific sets of institu-
tional forms, procedures or symbols; and it emphasizes how such practices are
diffused through organizational fields or across nations. They are interested, for
instance, in explaining the striking similarities in organizational form and prac-
tice that Education Ministries display throughout the world, regardless of differ-
ences in local conditions, or that firms display across industrial sectors whatever
the product they manufacture. Dobbin employs the approach to show how cul-
turally-constructed conceptions of the state and market conditioned 19th century
railroad policy in France and the United States.50 Meyer and Scott use it to ex-
plain the proliferation of training programs in American firms.51 Others apply it
to explain institutional isomorphism in East Asia and the relative facility with
which East Asian production techniques were diffused throughout the world.52
Fligstein takes this approach to explaining the diversification of American indus-
try, and Soysal uses it to explain contemporary immigration policy in Europe and
America.53

Three features of sociological institutionalism render it relatively distinctive in


the context of the other ‘new institutionalisms.’

First, the sociological institutionalists tend to define institutions much more


broadly than political scientists do to include, not just formal rules, procedures or
norms, but the symbol systems, cognitive scripts, and moral templates that pro-
vide the ‘frames of meaning’ guiding human action.54 Such a definition breaks
down the conceptual divide between ‘institutions’ and ‘culture.’ The two shade
Hall, Taylor: The Three New Institutionalisms 15

into each other. This has two important implications. First, it challenges the dis-
tinction that many political scientists like to draw between ‘institutional explana-
tions’ based on organizational structures and ‘cultural explanations’ based on an
understanding of culture as shared attitudes or values.55 Second, this approach
tends to redefine ‘culture’ itself as ‘institutions.’56 In this respect, it reflects a
‘cognitive turn’ within sociology itself away from formulations that associate cul-
ture exclusively with affective attitudes or values toward ones that see culture as
a network of routines, symbols or scripts providing templates for behavior.57

Second, the new institutionalists in sociology also have a distinctive understand-


ing of the relationship between institutions and individual action, which follows
the ‘cultural approach’ described above (pp. 5–6) but displays some characteristic
nuances. An older line of sociological analysis resolved the problem of specifying
the relationship between institutions and action by associating institutions with
‘roles’ to which prescriptive ‘norms of behavior’ were attached. In this view, in-
dividuals who have been socialized into particular institutional roles internalize
the norms associated with these roles, and in this way institutions are said to af-
fect behavior. We might think of this as the ‘normative dimension’ of institutional
impact. Although some continue to employ such conceptions, many sociological
institutionalists put a new emphasis on what we might think of as the ‘cognitive
dimension’ of institutional impact. That is to say, they emphasize the way in
which institutions influence behavior by providing the cognitive scripts, catego-
ries and models that are indispensable for action, not least because without them
the world and the behavior of others cannot be interpreted.58 Institutions influ-
ence behavior not simply by specifying what one should do but also by specify-
ing what one can imagine oneself doing in a given context. Here, one can see the
influence of social constructivism on the new institutionalism in sociology. In
many cases, institutions are said to provide the very terms through which mean-
ing is assigned in social life. It follows that institutions do not simply affect the
strategic calculations of individuals, as rational choice institutionalists contend,
but also their most basic preferences and very identity. The self-images and iden-
tities of social actors are said to be constituted from the institutional forms, im-
ages and signs provided by social life.59

Accordingly, many sociological institutionalists emphasize the highly-interactive


and mutually-constitutive character of the relationship between institutions and
individual action. When they act as a social convention specifies, individuals si-
multaneously constitute themselves as social actors, in the sense of engaging in
socially meaningful acts, and reinforce the convention to which they are adher-
ing. Central to this perspective is the notion that action is tightly bound up with
interpretation. Thus, the sociological institutionalists insist that, when faced with
a situation, the individual must find a way of recognizing it as well as of respond-
16 MPIFG Discussion Paper 96/6

ing to it, and the scripts or templates implicit in the institutional world provide
the means for accomplishing both of these tasks, often more or less simultane-
ously. The relationship between the individual and the institution, then, is built
on a kind of ‘practical reasoning’ whereby the individual works with and re-
works the available institutional templates to devise a course of action.60

None of this suggests that individuals are not purposive, goal-oriented or ra-
tional. However, sociological institutionalists emphasize that what an individual
will see as ‘rational action‘ is itself socially constituted, and they conceptualize
the goals toward which an actor is striving in much broader terms than others do.
If rational choice theorists often posit a world of individuals or organizations
seeking to maximize their material well-being, sociologists frequently posit a
world of individuals or organizations seeking to define and express their identity
in socially appropriate ways.

Finally, the new institutionalists in sociology also take a distinctive approach to


the problem of explaining how institutional practices originate and change. As
we have seen, many rational choice institutionalists explain the development of
an institution by reference to the efficiency with which it serves the material ends
of those who accept it. By contrast, sociological institutionalists argue that or-
ganizations often adopt a new institutional practice, not because it advances the
means-ends efficiency of the organization but because it enhances the social le-
gitimacy of the organization or its participants. In other words, organizations
embrace specific institutional forms or practices because the latter are widely val-
ued within a broader cultural environment. In some cases, these practices may ac-
tually be dysfunctional with regard to achieving the organization’s formal goals.
Campbell captures this perspective nicely by describing it as a ‘logic of social ap-
propriateness’ in contrast to a ‘logic of instrumentality.’61

Thus, in contrast to those who explain the diversification of American firms in the
1950s and 1960s as a functional response to economic or technological exigency,
Fligstein argues that managers embraced it because of the value that became as-
sociated with it in the many professional forums in which they participated and
the validation it offered for their broader roles and worldviews.62 Similarly,
Soysal argues that the policies toward immigrants adopted by many states were
pursued, not because they were most functional for the state, but because the
evolving conceptions of human rights promulgated by international regimes
made such policies seem appropriate and others illegitimate in the eyes of na-
tional authorities.63

Central to this approach, of course, is the question of what confers ‘legitimacy’ or


‘social appropriateness’ on some institutional arrangements but not others. Ulti-
mately, this is an issue about the sources of cultural authority. Some of the socio-
Hall, Taylor: The Three New Institutionalisms 17

logical institutionalists emphasize the way in which a modern state of expanding


regulatory scope imposes many practices on societal organizations by public fiat.
Others stress the way in which the growing professionalization of many spheres
of endeavor creates professional communities with the cultural authority to press
certain standards on their members.64 In other cases, common institutional prac-
tices are said to emerge from a more interactive process of discussion among the
actors in a given network – about shared problems, how to interpret them, and
how to solve them – taking place in a variety of forums that range from business
schools to international conclaves. Out of such interchanges, the actors are said to
develop shared cognitive maps, often embodying a sense of appropriate institu-
tional practices, which are then widely deployed. In these instances, the interac-
tive and creative dimensions of the process whereby institutions are socially con-
structed is most apparent.65 Some argue that we can even see such processes at
work on a transnational scale, where conventional concepts of modernity confer a
certain measure of authority on the practices of the most ‘developed‘ states and
exchanges under the aegis of international regimes encourage shared understand-
ings that carry common practices across national boundaries.66

5 Comparing Institutionalisms

In all their varieties, the ‘new institutionalisms’ significantly advance our under-
standing of the political world. However, the images they present of the political
world are by no means identical; and each displays characteristic strengths and
weaknesses. We consider these, first, with respect to the problem of specifying
the relationship between institutions and behavior.

Historical institutionalism has the most commodious conception of this relation-


ship. Analysts in this school commonly utilize both ‘calculus’ and ‘cultural’ ap-
proaches to this problem – in our view an important virtue, since we find both
perspectives plausible and important. However, eclecticism has its costs: histori-
cal institutionalism has devoted less attention than the other schools to develop-
ing a sophisticated understanding of exactly how institutions affect behavior, and
some of its works are less careful than they should be about specifying the precise
causal chain through which the institutions they identify as important are affect-
ing the behavior they are meant to explain. This is one respect in which historical
institutionalism might benefit from greater interchange among the schools.

Rational choice institutionalism, by contrast, has developed a more precise con-


ception of the relationship between institutions and behavior and a highly gen-
18 MPIFG Discussion Paper 96/6

eralizable set of concepts that lend themselves to systematic theory-building.


However, these widely-vaunted microfoundations rest on a relatively simplistic
image of human motivation, which may miss many of its important dimen-
sions.67 Defenders of the approach are inclined to compare it to a set of reduced-
form equations, properly judged not on the accuracy of their assumptions but on
the predictive power of their models.68 But this is treacherous ground, since the
predictions generated by such models are often sensitive to small changes in as-
sumptions about pay-off matrices, preference structures and the like, which are
frequently arbitrary or unsupported by data.69 The usefulness of the approach is
also limited by the degree to which it specifies the preferences or goals of the ac-
tors exogenously to the analysis, especially in empirical cases where these under-
lying preferences are multifaceted, ambiguous or difficult to specify ex ante.
Since instrumental behavior is a major component of politics, however, rational
choice institutionalism has made major contributions, notably by highlighting
key aspects of politics that are often underappreciated by other perspectives and
providing tools for analyzing them. Members of this school emphasize that po-
litical action involves the management of uncertainty, long one of the most cen-
tral and neglected features of politics; and they demonstrate the importance that
flows of information have for power relations and political outcomes.

Perhaps most important, they draw our attention to the role that strategic inter-
action between actors plays in the determination of political outcomes. This rep-
resents a major advance on traditional approaches that explain political outcomes
largely in terms of the force that structural variables, such as level of socioeco-
nomic development, educational attainment or material discontent, are said to
exercise directly over individual behavior. With it, rational choice analysts can in-
corporate into their analyses a much more extensive appreciation for the role that
human intentionality plays in the determination of political outcomes, in the
form of strategic calculation, integrated with a role for structural variables under-
stood primarily in terms of institutions. The difference between the two ap-
proaches is epitomized in the movement from models in which causality is repre-
sented by the coefficients of structural variables in regression equations toward
game-theoretic models of political processes. The drawback, of course, is that this
advance comes at the cost of conceptualizing intentionality in terms of a rela-
tively thin theory of human rationality.

Anyone who has waited at a traffic light when no one else was around, however,
has to admit that there are dimensions to the relationship between institutions
and action that may not be highly instrumental or well-modeled by rational
choice theories. Sociological institutionalists are often better placed to elucidate
these dimensions. On the one hand, their theories specify ways in which institu-
tions can affect the underlying preferences or identities of actors that rational
Hall, Taylor: The Three New Institutionalisms 19

choice institutionalists must take as given. On the other hand, they tell us that
even a highly instrumental actor may be choosing strategies (and rivals) from cul-
turally-specific repertoires, thereby identifying additional respects in which the
institutional environment may affect the strategies that actors choose. There is
some sense in which the sociologists capture aspects of institutional impact that
may be the indispensable antecedents to instrumental action.70

Turning to the second of our organizing issues, there are also distinctive
strengths and weaknesses in the approaches taken by the three schools to the
problem of explaining how institutions originate and change.

Rational choice institutionalists have produced the most elegant accounts of insti-
tutional origins, turning primarily on the functions that these institutions per-
form and the benefits they provide. In our view, this approach has real strength
for explaining why existing institutions continue to exist, since the persistence of
an institution often depends upon the benefits it can deliver. However, several
features of the approach severely limit its adequacy as a framework for explain-
ing the origins of institutions.

First, the approach of rational choice institutionalism is often highly ‘func-


tionalist.’ That is to say, it explains the origins of an institution largely in terms of
the effects that follow from its existence. Although such effects may contribute to
the persistence of an institution, the problem of explaining persistence should not
be confused with the problem of explaining an institution’s origins. Because un-
intended consequences are ubiquitous in the social world, one cannot safely de-
duce origins from consequences.71 Moreover, this approach often leaves us with-
out an explanation for the many inefficiencies that institutions display, and it
may overstate the efficiency that some do display.72

Second, this approach is largely ‘intentionalist.’ In other words, it tends to assume


that the process of institutional creation is a highly purposive one, largely under
the control of actors who correctly perceive the effects of the institutions they es-
tablish and create them precisely in order to secure these effects. Although there
is undoubtedly a purposive element to institutional creation, such analyses often
entail heroic assumptions about the prescience of historical actors and their ca-
pacity to control the course of events. In some cases, they also impute overly-
simple intentions to historical actors who, on closer inspection, may be seen to be
operating from a much more complex set of motivations.73

Third, many such analyses are highly ‘voluntarist.’ That is to say, as Bates has ar-
gued, they tend to view institutional creation as a quasi-contractual process
marked by voluntary agreement among relatively equal and independent actors
– much as one might find in a ‘state of nature.’74 Although this may accurately
20 MPIFG Discussion Paper 96/6

depict some cases, there are many others in which such an approach understates
the degree to which asymmetries of power vest some actors with more influence
than others over the process of institutional creation.75 Finally, the ‘equilibrium’
character of the rational choice approach to institutions embroils such analysts in
a contradiction. One implication of this approach is that the starting-point from
which institutions are to be created is itself likely to reflect a Nash equilibrium.
Thus, it is not obvious why the actors would agree to a change in existing institu-
tions. Paradoxically, the efforts of Shepsle and others to show that institutions are
stable, by invoking the uncertainty that surrounds institutional change, render
the problem of explaining why institutions change even more perplexing.76 At a
minimum, this approach needs a more robust theory of dynamic equilibria.

These considerations suggest that, although rational choice institutionalism has


great potential for explaining why institutions continue to persist, the explana-
tion it offers for institutional genesis probably applies well only to a limited
number of settings. Specifically, it offers the greatest analytical leverage in set-
tings where consensus among actors accustomed to strategic action and of
roughly equal standing is necessary to secure institutional change, as in some
legislatures or international arenas. Alternatively, it may be applicable to settings
where intense competition among organizational forms selects for those with
some kind of efficiency that is clearly specifiable ex ante, as in some settings of
market competition.77

By contrast, historical and sociological institutionalists approach the problem of


explaining how institutions originate and change quite differently. Both begin by
insisting that new institutions are created or adopted in a world already replete
with institutions. This may seem a simple point, but much follows from it.

Sociological institutionalists use it to explore the way in which existing institu-


tions structure the field of vision of those contemplating institutional reform.
Thus, they focus attention on the processes whereby those developing new insti-
tutions ‘borrow’ from the existing world of institutional templates. This approach
usefully emphasizes the way in which the existing institutional world circum-
scribes the range of institutional creation. The sociological institutionalists also
develop a more expansive conception of why a particular institution might be
chosen, which goes beyond considerations of efficiency toward an appreciation
for the role that collective processes of interpretation and concerns for social le-
gitimacy play in the process. Among other things, such an approach goes a long
way toward explaining the presence of many apparent inefficiencies in social and
political institutions.78
Hall, Taylor: The Three New Institutionalisms 21

From the perspective of political science, however, the approach that sociological
institutionalism takes to such processes often seems curiously bloodless. That is
to say, it can miss the extent to which processes of institutional creation or reform
entail a clash of power among actors with competing interests.79 After all, many
actors, both inside and outside an organization, have deep stakes in whether that
firm or government adopts new institutional practices, and reform initiatives
often provoke power struggles among these actors, which an emphasis on proc-
esses of diffusion can neglect. In some cases, the new institutionalists in sociology
seem so focused on macro-level processes that the actors involved in these proc-
esses seem to drop from sight and the result begins to look like ‘action without
agents.’ In general, the approach as a whole might benefit from more attention to
the way in which frames of meaning, scripts and symbols emerge not only from
processes of interpretation but also from processes of contention.80

Historical institutionalists use the same starting-point, namely a world replete


with institutions, to direct our attention to the way in which the power relations
instantiated in existing institutions give some actors or interests more power than
others over the creation of new institutions.81 In this respect, they join with the
rational choice institutionalists to build on the famous point made by an earlier
generation of analysts that ‘organization is the mobilization of bias.’82 However,
to this emphasis, they marry a conception of path dependence that also recog-
nizes the importance of existing institutional templates to processes of institu-
tional creation and reform.

If rational choice accounts of the origin of institutions are dominated by deduc-


tion, those of the historical institutionalists often seem to depend heavily on in-
duction. Typically, they scour the historical record for evidence about why the
historical actors behaved as they did. This neo-Weberian focus on the meanings
that historical actors attribute to their own actions greatly enhances the realism of
the analyses produced by historical institutionalists, and it allows them to dis-
criminate among competing explanations when the deductive calculus associated
with rational actors specifies more than one equilibrium outcome. As a result
they have produced some startling revisions to our conventional understandings
about the origin of such institutions as Swedish corporatism.83 However, this
emphasis on induction has been a weakness as well as a strength: historical insti-
tutionalists have been slower than others to aggregate their findings into system-
atic theories about the general processes involved in institutional creation and
change.
22 MPIFG Discussion Paper 96/6

6 Conclusion

In sum, political science today is confronted with not one but three ‘new institu-
tionalisms.’ Moreover, it is striking how distant these schools of thought have
remained from each other. Each has been assiduously burnishing its own para-
digm. What is the way forward? Many argue for the wholehearted embrace of
one of these approaches at the expense of the others. However, the thrust of this
essay has been to suggest that the time has come for greater interchange among
them. At a minimum, it suggests that a better acquaintance with the other schools
would lead the partisans of each toward a more sophisticated appreciation for the
underlying issues still to be resolved within their own paradigm.

Can this interchange go farther? Might each of the schools borrow or adapt some
of the insights developed by the others? There would be limits to any such inte-
gration. When confronting each other on the highly theoretical terrain of first
principles, the most extreme proponents of each approach take very different
positions on such fundamental issues as whether the identities of the actors can
be given exogenously to the institutional analysis and whether it makes sense to
assume a homogenous kind of rational or strategic action across cultural settings.

However, we favor taking this interchange as far as possible, most fundamentally


because each of these literatures seems to reveal different and genuine dimen-
sions of human behavior and of the effects institutions can have on behavior.
None of these literatures appears to be wrong-headed or substantially untrue.
More often, each seems to be providing a partial account of the forces at work in a
given situation or capturing different dimensions of the human action and insti-
tutional impact present there.

For instance, an actor’s behavior may be influenced both by strategic calculation


about the likely strategies of others and by reference to a familiar set of moral or
cognitive templates, each of which may depend on the configuration of existing
institutions. Consider the case of French workers contemplating adherence to an
incomes policy during the 1950s. On the one hand, the divided structure of the
French labor movement discouraged a strategy of adherence because it was con-
ducive to free-riding. On the other hand, the syndicalist ideologies of many
French unions also militated against cooperation in such an endeavor.84 It is pos-
sible that there were two respects in which the institutions of the French labor
movement were influencing behavior at this time, each modeled more effectively
by a different school of thought.

Moreover, if the most extreme assumptions of each school’s theoretical position


are relaxed, they share a great deal of common analytical ground on which the
Hall, Taylor: The Three New Institutionalisms 23

insights of one approach might be used to supplement or strengthen those of an-


other. For instance, both the ‘calculus’ and ‘cultural’ approaches to the relation-
ship between institutions and action observe that institutions affect action by
structuring expectations about what others will do, even if they model the
sources of those expectations slightly differently. In one case, those expectations
are said to be shaped by what should seem instrumentally viable to the other ac-
tor; in the other they are said to be shaped by what should seem socially appro-
priate to the other actor. There is room for a useful dialogue here. Similarly, it
would not be difficult for proponents of the calculus and cultural approaches to
acknowledge that a good deal of behavior is goal-oriented or strategic but that
the range of options canvassed by a strategic actor is likely to be circumscribed by
a culturally-specific sense of appropriate action.

A number of analysts have already moved some distance in this direction in such
a way as to suggest that considerable promise may lie in such syntheses. In what
might otherwise be a conventional rational choice analysis of how organizations
monitor and enforce behavior among their employees, for instance, Kreps ex-
tends the argument to encompass ‘corporate culture,’ understood as a set of col-
lective templates for action. He argues that such ‘cultures’ can be an efficient
supplement to the traditional monitoring and enforcement mechanisms of an or-
ganization, especially when the latter cannot readily specify appropriate behavior
for all contingencies.85

Other rational choice analysts have begun to incorporate ‘culture’ or ‘beliefs’ into
their work to explain why actors move toward one outcome when a conventional
analysis specifies many possible equilibrium outcomes. Garrett and Weingast, for
example, argue that the norms or ideas fostered by a particular institutional envi-
ronment often provide the ‘focal points’ that allow rational actors to converge on
one among many possible equilibria.86 In an especially intriguing analysis of
games with multiple equilibria, Scharpf shows how behavior might be deter-
mined jointly by both the ‘decision rules’ that represent the incentives institutions
provide to the actors as rational calculators and the ‘decision styles’ of those ac-
tors, which can be interpreted to mean the beliefs about appropriate behavior
that cultural analysts emphasize. To take only one example, the latter may spec-
ify whether the actor attaches greater value to relative or absolute gains when the
payoff matrix specifies a choice between the two.87 Similarly, Bates and Weingast
argue that strategic interactions are signalling games, whose meanings and out-
comes can be specified only if we understand the cultural context that assigns
meaning to specific symbols; and they go farther to suggest that the object of
many kinds of strategic interaction may be precisely to affect such beliefs.88
24 MPIFG Discussion Paper 96/6

Historical institutionalism stands in an especially pivotal position. Many of the


arguments recently produced by this school could readily be translated into ra-
tional choice terms, while others display clear openings toward the new institu-
tionalism in sociology.89 The best of these analyses already effect something of in-
tegration, for instance, by showing how historical actors select new institutions
for instrumental purposes, much as a rational choice analysis would predict, but
draw them from a menu of alternatives that is made historically available
through the mechanisms specified by sociological institutionalism.90 As noted
above, others have gone farther to suggest that strategic responses to a particular
institutional environment may eventually give rise to worldviews and organiza-
tional practices that continue to condition action even after the initial institutional
environment has changed.91

Let us be clear: we are not arguing that a crude synthesis of the positions devel-
oped by each of these schools is immediately practicable or even necessarily de-
sirable. After all, it is precisely because the implicit debate among them has been
enlightening that we have tried to make it more explicit here, and there is much
to be said for tenacious debate. Our main point is that, after some years in which
these schools of thought have incubated in relative isolation from each other, the
time has come for a more open and extensive interchange among them. There is
ample evidence that we can learn from all of these schools of thought and that
each has something to learn from the others.
Hall, Taylor: The Three New Institutionalisms 25

7 Notes

An earlier version of this paper was presented at the 1994 Annual Meeting of the Ameri-
can Political Science Association and at a Conference on ‘What is Institutionalism Now?’
at the University of Maryland, October 1994. We would like to acknowledge the hospi-
tality and stimulation that W. Richard Scott, the Stanford Center for Organizations Re-
search, and the Center for Advanced Study in the Behavioral Sciences provided while
the preliminary work for this paper was being done, and we are grateful to Paul Pierson
for many helpful discussions about these issues. For written comments on this earlier
draft, we are grateful to Robert Bates, Paul DiMaggio, Frank Dobbin, James Ennis, Bar-
bara Geddes, Peter Gourevitch, Ian Lustick, Cathie Jo Martin, Lisa Martin, Paul Pierson,
Mark Pollack, Bo Rothstein, Kenneth Shepsle, Rogers Smith, Marc Smyrl, Barry Wein-
gast, and Deborah Yashar.

1 In principle, we might also identify a fourth such school, the ‘new institutionalism’
in economics. However, it and rational choice institutionalism overlap heavily and
so we treat them together in this brief review. A more extended treatment might ob-
serve that RCI puts more emphasis on strategic interaction, while the new institu-
tionalism in economics puts more stress on property rights, rents, and competitive
selection mechanisms. Cf. Thrainn Eggertsson, Economic Behavior and Institutions
(Cambridge, Cambridge University Press, 1990) and Louis Putterman (ed.), The Eco-
nomic Nature of the Firm (Cambridge, Cambridge University Press, 1986).
2 We borrow the term ‘historical institutionalism’ from Sven Steinmo et al., Structuring
Politics: Historical Institutionalism in Comparative Analysis (New York, Cambridge
University Press, 1992).
3 Of necessity, this is a highly concise summary of many complex developments. For
amplification, see: Ronald Chilcote, Theories of Comparative Politics (Boulder,
Westview, 1981) and James A. Bill and Robert L. Hardgrave Jr., Comparative Politics
(Washington, University Press of America, 1981).
4 Cf. Harry Eckstein and David Apter (eds.), Comparative Politics (Glencoe, Free Press,
1963).
5 For an especially influential and integrated treatment, see Gabriel Almond and G.
Bingham Powell, Jr., Comparative Politics: A Developmental Approach (Boston, Little
Brown, 1956).
6 Cf. Robin Blackburn (ed.), Ideology and Social Science (London, Fontana, 1972), ch. 11;
Fred Block, Revising State Theory (Philadelphia, Temple University Press, 1987), and
Martin Carnoy, The State and Political Theory (Princeton, Princeton University Press,
1984).
7 Cf. Peter Evans et al. (eds.), Bringing the State Back In (New York, Cambridge Uni-
versity Press, 1985), Stephen Krasner, Defending the National Interest (Princeton, Prin-
ceton University Press, 1980), and Peter Katzenstein (ed.), Between Power and Plenty
(Madison, University of Wisconsin Press, 1978).
8 There was a significant point of tangency with the literature on neo-corporatism.
See: John Zysman, Governments, Markets and Growth (Berkeley, University of Cali-
fornia Press, 1983), Philippe Schmitter and Gerhard Lehmbruch (eds.), Patterns of
Corporatist Policy-Making (Beverly Hills, Sage, 1982), and Peter A. Hall, Governing the
Economy: The Politics of State Intervention in Britain and France (Oxford, Polity, 1986).
26 MPIFG Discussion Paper 96/6

9 Cf. Steinmo et al. (eds.), op. cit. in note 2, and R. Kent Weaver and Bert A. Rockman
(eds.), Do Institutions Matter? (Washington, Brookings, 1993).
10 Cf. John Goldthorpe (ed.), Order and Conflict in Contemporary Capitalism (New York,
Cambridge University Press, 1984), David Soskice, ‘Wage Determination: The
Changing Role of Institutions in Advanced Industrialized Countries,’ Oxford Review
of Economic Policy 6,4 (1990), pp. 36–61, Fritz W. Scharpf, Crisis and Choice in Social
Democracy (Ithaca, Cornell University Press, 1992).
11 See Kathleen Thelen and Sven Steinmo, ‘Historical Institutionalism in Comparative
Politics’ in Steinmo et al. (eds.), Structuring Politics, p. 2 ff; Hall, Governing the Econ-
omy, p. 19. For a thoughtful example of a broader view, see: John Ikenberry,
‘Conclusion: An Institutional Approach to American Foreign Policy’ in John Iken-
berry et al. (eds.), The State and American Foreign Policy (Ithaca, Cornell University
Press, 1988), p. 226.
12 For an excellent overview from which our analysis has benefited, see: G. John Iken-
berry, ‘History’s Heavy Hand: Institutions and the Politics of the State.’ Paper pre-
sented to a Conference on ‘What is Institutionalism Now?,’ University of Maryland,
October 1994.
13 For an especially trenchant characterization of this position, see James March and
Johan P. Olsen, Rediscovering Institutions: The Organizational Basis of Politics (New
York, Free Press, 1989).
14 These can also be seen as responses to one of the most important dimensions of the
structure-agency problem, namely, the problem of explaining how an institution can
be said to structure human action, in some determinative sense, so as to produce a
regularized pattern of behavior, even though the existence of the institution itself
usually depends on the presence of these patterns of behavior and thus on the will-
ingness of the actors to behave in certain ways. The problem is to capture simultane-
ously the voluntary and determinative character of institutions. For a more general
exploration of such problems, see Anthony Giddens, Central Problems in Social Theory
(London, Macmillan, 1978).
15 For a radical statement of this view, see: Randall L. Calvert, ‘The Rational Choice
Theory of Social Institutions’ in Jeffrey S. Banks and Eric A. Hanushek (eds.), Modern
Political Economy (New York, Cambridge University Press, 1995), pp. 216–266.
16 To this Kenneth Shepsle has added the observation that actors will hesitate to
change the institutional rules because, although reform might allow them to realize
an immediate gain on the issue-at-hand, they face great uncertainty about the impact
of the new rules on decisions not yet foreseen. See: Kenneth A. Shepsle, ‘Institutional
Equilibrium and Equilibrium Institions’ in Herbert F. Weisberg (ed.), Political Science:
The Science of Politics (New York, Agathon, 1986), pp. 51–81.
17 For a radical critique which begins from this kind of point but moves well beyond it,
see: Robert Graftstein, Institutional Realism: Social and Political Constraints on Rational
Actors (New Haven, Yale University Press, 1992).
18 Ellen Immergut, Health Politics: Interests and Institutions in Western Europe (New York,
Cambridge University Press, 1992).
19 Victoria C. Hattam, Labor Visions and State Power: The Origins of Business Unionism in
the United States (Princeton, Princeton University Press, 1993).
20 Cf. Steven Lukes, Power: A Radical View (London, Macmillan, 1972) and John
Gaventa, Power and Powerlessness: Quiescence and Rebellion in an Appalachian Valley
(Urbana, University of Illinois Press, 1980).
Hall, Taylor: The Three New Institutionalisms 27

21 Sven Steinmo, Taxation and Democracy: Swedish, British and American Approaches to Fi-
nancing the Modern State (New Haven, Yale University Press, 1993).
22 Margaret Weir, ‘Ideas and the Politics of Bounded Innovation’ in Steinmo et al., op.
cit. in note 2, pp. 188–216.
23 Cf. David Collier and Ruth Collier, Shaping the Political Arena (Princeton, Princeton
University Press, 1991); M. Downing, The Military Revolution and Political Change:
Origins of Democracy and Autocracy in Early Modern Europe (Princeton, Princeton Uni-
versity Press, 1992); and Stephen Krasner, ‘Sovereignty: An Institutional Perspec-
tive,’ Comparative Political Studies 21 (1988), pp. 66–94.
24 See Margaret Weir and Theda Skocpol, ‘State Structures and the Possibilities for
‘Keynesian’ Responses to the Great Depression in Sweden, Britain and the United
States’ in Peter B. Evans et al., Bringing the State Back In (Cambridge, Cambridge
University Press, 1985), pp. 107–163.
25 Cf. Paul Pierson, Dismantling the Welfare State? (Cambridge, Cambridge University
Press, 1994) and ‘When Effect Becomes Cause: Policy Feedback and Political
Change,’ World Politics 45, 4 (July 1993), pp. 595–628; Jane Jenson, ‘Paradigms and
Political Discourse: Protective Legislation in France and the United States Before
1914,’ Canadian Journal of Political Science 22, 2 (June 1989), pp. 235–258; Ira Katznel-
son, City Trenches: Urban Politics and the Patterning of Class in the United States (New
York, Pantheon Books, 1981).
26 Cf. James March and Johan Olsen, ‘The New Institutionalism: Organizational Fac-
tors in Political Life,’ American Political Science Review 78 (Sept. 1984), pp. 734–749
and Douglass C. North, Institutions, Institutional Change and Economic Performance
(New York, Cambridge University Press, 1990).
27 Cf. Peter A. Gourevitch, Politics In Hard Times (Ithaca, Cornell University Press,
1986); David Collier and Ruth Collier, op. cit. in note 23, and Stephen Krasner,
‘Approaches to the State: Alternative Conceptions and Historical Dynamics,’ Com-
parative Politics 16, 2 (January, 1984), pp. 223–46.
28 This issue deserves fuller examination than scholars have yet given it. For one view,
see: Theda Skocpol, States and Social Revolutions (New York, Cambridge University
Press, 1979).
29 See: Judith Goldstein, ‘Ideas, Institutions and American Trade Policy,’ International
Organization 42, 1 (1988), pp. 179–217: Margaret Weir, ‘Ideas and Politics: The Ac-
ceptance of Keynesianism in Britain and the United States’ in Peter A. Hall (ed.), The
Political Power of Economic Ideas (Princeton, Princeton University Press, 1989), pp. 53–
86; Kathryn S. Sikkink, Ideas and Institutions: Developmentalism in Brazil and Argentina
(Ithaca, Cornell University Press, 1991
30 The seminal work is: William Riker, ‘Implications from the Disequilibrium of Ma-
jority Rule for the Study of Institutions,’ American Political Science Review 74 (1980),
pp. 432–447. See also: Richard McCelvey, ‘Intransitivities in Multidimensional Vot-
ing Models and Some Implications for Agenda Control,’ Journal of Economic Theory
12 (1976), pp. 472–482; and John Ferejohn and Morris Fiorina, ‘Purposive Models of
Legislative Behavior,’ American Economic Review, Papers and Proceedings 65 (1975), pp.
407–415.
31 Kenneth A. Shepsle, ‘Institutional Equilibrium and Equilibrium Institutions’ in Her-
bert F. Weisberg (ed.), Political Science: The Science of Politics (New York, Agathon,
1986), pp. 51–81; and Shepsle, ‘Studying Institutions: Some Lessons from the Ra-
tional Choice Approach,’ Journal of Theoretical Politics 1, 2 (1989), pp. 131–147.
28 MPIFG Discussion Paper 96/6

32 Cf. Barry Weingast and William Marshall, ‘The Industrial Organization of Con-
gress,’ Journal of Political Economy 96, 1 (1988), pp. 132–163.
33 Two of the seminal articles are: Terry Moe, ‘The New Economics of Organization,’
American Journal of Political Science 28 (1984), pp. 739–777; and Weingast and Mar-
shall, ‘The Industrial Organization of Congress,’ op. cit. in note 32.
34 Oliver Williamson, Markets and Hierarchies (New York, Free Press, 1975) and Wil-
liamson, The Economic Institutions of Capitalism (New York, Free Press, 1985).
35 Cf. Douglass C. North and Paul Thomas, The Rise of the Western World (New York,
Cambridge University Press, 1973).
36 Paul Milgrom and John Roberts, Economics, Organization and Management (New York,
Prentice-Hall, 1992); John W. Pratt and Richard Zeckhauser, Principals and Agents
(Boston, Harvard Business School Press, 1991).
37 Matthew D. McCubbins and Terry Sullivan (eds.), Congress: Structure and Policy
(New York, Cambridge University Press, 1987) and Legislative Studies Quarterly (May
1994).
38 Gary Cox and Mathew D. McCubbins, Legislative Leviathan (Berkeley, University of
California Press, 1987); John Ferejohn, ‘Law, Legislation and Positive Political The-
ory’ in Banks and Hanushek, Modern Political Economy: Old Topics, New Directions
(New York, Cambridge University Press), pp. 191–215; Kenneth A. Shepsle and
Barry R. Weingast, ‘Positive Theories of Congressional Institutions,’ Legislative Stud-
ies Quarterly (May 1994); Terry Moe, ‘An Assessment of the Positive Theory of Con-
gressional Dominance,’ Legislative Studies Quarterly 12, 4 (1987), pp. 475–520; Mat-
thew McCubbins and Thomas Schwartz, ‘Congressional Oversight Overlooked: Po-
lice Patrols versus Fire Alarms,’ American Journal of Political Science 28 (February
1984), pp. 165–179.
39 Michael Laver and Kenneth Shepsle, ‘Coalitions and Cabinet Government,’ American
Political Science Review 84 (1990), pp. 843–890; Douglass C. North and Barry Wein-
gast, ‘Constitutions and Credible Commitments: The Evolution of Institutions Gov-
erning Public Choice in 17th Century England,’ Journal of Economic History 49
(December 1989), pp. 803–832; Barry Weingast, ‘Institutionalizing Trust: The Political
and Economic Roots of Ethnic and Regional Conflict.’ Paper presented to a confer-
ence on ‘What is Institutionalism Now?,’ University of Maryland, 1994.
40 Adam Przeworski, Democracy and the Market (Cambridge, Cambridge University
Press, 1991); Barbara Geddes, Politicians’ Dilemma (Berkeley, University of California
Press, 1994); Gary Marks, ‘Rational Sources of Chaos in Democratic Transitions,’
American Behavioral Scientist 33 4/5 (1992), pp. 397–421; Youssef Cohen, Radicals, Re-
formers and Reactionaries (Chicago, Chicago University Press, 1994); James deNardo,
Power in Numbers (Princeton, Princeton University Press, 1985).
41 George Tsebelis, ‘The Power of the European Parliament as a Conditional Agenda
Setter,’ American Political Science Review 88, 1 (1994), pp. 795–815; Mark Pollack,
‘Obedient Servant or Runaway Eurocracy?,’ Harvard Center for European Studies
Working Paper, 1995; Lisa Martin, ‘The Influence of National Parliaments on Euro-
pean Integration,’ Harvard Center for International Affairs Working Paper, 1994.
42 Robert O. Keohane and Lisa L. Martin, ‘Delegation to International Organizations.’
Paper presented to a conference on ‘What is Institutionalism Now?,’ University of
Maryland, 1994; Lisa Martin, ‘Interests, Power and Multilateralism,’ International Or-
ganization 46, 4 (Autumn 1992), pp. 765–792; Kenneth A. Oye (ed.), Cooperation Under
Anarchy (Princeton, Princeton University Press, 1993); Stephen Krasner, ‘Global
Hall, Taylor: The Three New Institutionalisms 29

Communications and National Power: Life on the Pareto Frontier,’ World Politics 43
(April 1991), pp. 336–366.
43 Cf. Kenneth Shepsle and Barry Weingast, ‘The Institutional Foundations of Commit-
tee Power,’ American Political Science Review 81 (March 1987), pp. 85–104; cf. Jon El-
ster and Aanund Hylland (eds.), Foundations of Social Choice Theory (Cambridge,
Cambridge University Press, 1986).
44 Cf. G. Hardin, ‘The Tragedy of the Commons,’ Science 162 (1968), pp. 1243–1248;
Russell Hardin, Collective Action (Baltimore, Johns Hopkins Press, 1982); Elinor Os-
trom, Governing the Commons (New York, Cambridge University Press, 1990).
45 As one might expect, analyses focused on legislatures tend to emphasize the impor-
tance of voluntary agreement, while analyses focused on economic institutions put
more emphasis on competitive selection.
46 Cf. Williamson, Market and Hierarchies (New York, Free Press, 1975), and Milgrom
and Roberts, Economics, Organization and Management; and Paul Milgrom and John
Roberts, ‘Bargaining Costs, Influence Costs and the Organization of Economic Ac-
tivity’ in James Alt and Kenneth Shepsle (eds.), Perspectives on Positive Political Econ-
omy (Cambridge, Cambridge University Press, 1990) pp. 57–89.
47 Keith Krehbiel, Information and Legislative Organization (Ann Arbor, University of
Michigan Press, 1991); cf. Kenneth Shepsle and Barry Weingast, Positive Theories of
Congressional Institutions (Ann Arbor, University of Michigan Press, 1995).
48 For amplification of these points, see: Frank Dobbin, ‘Cultural Models of Organiza-
tion: The Social Construction of Rational Organizing Principles’ in Diana Crane
(ed.), The Sociology of Culture (Oxford, Blackwell, 1994), pp. 117–153.
49 The pioneering work was done by Stanford sociologists. See: John W. Meyer and
Brian Rowan, ‘Institutionalized Organizations: Formal Structure as Myth and Cere-
mony,’ American Journal of Sociology 83 (1977), pp. 340–363; John W. Meyer and W. R.
Scott, Organizational Environments: Ritual and Rationality (Beverly Hills, Sage, 1983);
and Paul J. DiMaggio and Walter W. Powell, ‘Introduction’ in Powell and DiMaggio
(eds.), The New Institutionalism in Organizational Analysis (Chicago, University of Chi-
cago Press, 1991), pp. 1–40.
50 Frank Dobbin, Forging Industrial Policy (Cambridge, Cambridge University Press,
1994).
51 W. Richard Scott and John W. Meyer et al., Institutional Environments and Organiza-
tions (Thousand Oaks, CA, Sage, 1994), chs. 11, 12.
52 Marco Orru et al., ‘Organizational Isomorphism in East Asia’ in Powell and DiMag-
gio (eds.), pp. 361–389 and Robert E. Cole, Strategies for Industry: Small-Group Activi-
ties in American, Japanese and Swedish Industry (Berkeley, University of California
Press, 1989).
53 Neil Fligstein, The Transformation of Corporate Control (Cambridge, Harvard Univer-
sity Press, 1990); Yasemin N. Soysal, Limits of Citizenship (Chicago, University of Chi-
cago Press, 1994).
54 Cf. John L. Campbell, ‘Institutional Analysis and the Role of Ideas in Political Econ-
omy.’ Paper presented to the Seminar on the State and Capitalism since 1800, Har-
vard University, 1995; and W. Richard Scott, ‘Institutions and Organizations: To-
wards a Theoretical Synthesis’ in W. Richard Scott, John W. Meyer et al., Institutional
Environments and Organizations: Structural Complexity and Individualism (Thousand
Oaks CA, Sage, 1994), pp. 55–80.
55 Cf. Gabriel Almond and Sidney Verba, The Civic Culture (Boston, Little Brown, 1963)
and Peter A. Hall, Governing the Economy (Oxford, Polity, 1986), ch. 1.
30 MPIFG Discussion Paper 96/6

56 Cf. Lynne Zucker, ‘The Role of Institutionalization in Cultural Persistence’ in Powell


and DiMaggio (eds.), op. cit. in note 49, pp. 83–107; John Meyer et al., ‘Ontology and
Rationalization in the Western Cultural Account’ in John W. Meyer, W. Richard
Scott et al., op. cit. in note 54.
57 Cf. Ann Swidler, ‘Culture in Action: Symbols and Strategies,’ American Sociological
Review 51 (1986), pp. 273–286 and James March and Johan P. Olsen, Rediscovering In-
stitutions (New York, The Free Press, 1989), ch. 3.
58 Cf. DiMaggio and Powell, ‘Introduction’, op. cit. in note 49.
59 See the classic work by Peter Berger and Thomas Luckmann, The Social Construction
of Reality (New York, Anchor, 1966) and the more recent application to political sci-
ence by Alexander Wendt, ‘The Agent-Structure Problem in International Relations
Theory,’ International Organization 41, 3 (Summer 1987), pp. 335–370.
60 Cf. Paul J. DiMaggio and Walter W. Powell, ‘Introduction’ in Powell and DiMaggio,
op. cit. in note 49, pp. 22–24 and the essays by Lynne G. Zucker and Ronald Jepper-
son in that volume.
61 C.f. Campbell, ‘Institutional Analysis and the Role of Ideas in Political Economy,’
op. cit. in note 54, and James G. March and Johan P. Olsen, Rediscovering Institutions
(New York, Free Press, 1989), ch. 2.
62 Fligstein, op. cit. in note 53.
63 Soysal, op. cit. in note 53.
64 Cf. Paul J. DiMaggio and Walter W. Powell, ‘The Iron Cage Revisited: Institutional
Isomorphism and Collective Rationality’ and Walter W. Powell, ‘Expanding the
Scope of Institutional Analysis’ in Powell and DiMaggio, op. cit. in note 49, chs. 3
and 8.
65 On this point we are indebted to the insightful analysis of John L. Campbell, op. cit.
in note 54, p. 11.
66 See John W. Meyer et al., ‘Ontology and Rationalization,’ John W. Meyer,
‘Rationalized Environments’ and David Strang and John W. Meyer, ‘Institutional
Conditions for Diffusion’ in Scott, Meyer et al., op. cit in note 54, chs. 1, 2 and 5.
67 For more extended discussions, see: Karen S. Cook and Margaret Levi (eds.), The
Limits of Rationality (Chicago, University of Chicago Press, 1990) and Jane Mans-
bridge (ed.), Beyond Self-Interest (Chicago, University of Chicago Press, 1990).
68 We are grateful for Kenneth Shepsle for drawing this to our attention. Cf. Milton
Friedman, ‘The Methodology of Positive Economics,’ in Essays in Positive Economics
(Chicago, University of Chicago Press, 1953).
69 The problem is magnified by the fact that many equilibrium solutions may be pres-
ent in a given situation, as the ‘folk theorem’ suggests. More generally, see: Donald
P. Green and Ian Shapiro, Pathologies of Rational Choice Theory (New Haven, Yale
University Press, 1994).
70 Cf. James Johnson, ‘Symbolic Dimensions of Social Order.’ Paper presented to a con-
ference on ‘What is Institutionalism Now?,’ University of Maryland, 1994.
71 This is a point made in Robert Bates, ‘Contra-Contractarianism: Some Reflections on
the New Institutionalism,’ Politics and Society 16 (1987), pp. 387–401.
72 For a powerful effort to tackle this problem without entirely foresaking rational
choice assumptions, see: Terry Moe, ‘The Politics of Structural Choice: Toward a
Theory of Public Bureaucracy’ in Oliver Williamson (ed.), Organization Theory from
Chester Barnard to the Present and Beyond (New York, Oxford University Press, 1990),
pp. 116–153.
Hall, Taylor: The Three New Institutionalisms 31

73 For one example, see the otherwise valuable analysis in: Douglass C. North and
Barry Weingast, op. cit in note 39. Similarly, it may be that many rational choice
analyses assume too readily that the presence of collective action problems will gen-
erate an automatic ‘demand’ for new institutions. For correctives, see: Bates, ‘Contra-
Contractarianism’, op. cit. in note 71, and Jack Knight, Institutions and Social Conflict
(New York, Cambridge University Press, 1992).
74 See Bates, op. cit. in note 71, and Robert Grafstein, Institutional Realism: Social and Po-
litical Constraints on Rational Actors (New Haven, Yale University Press, 1992), ch. 3.
75 For an astute analysis that attempts to build an appreciation for asymmetries of
power into a rational choice theory of institutional creation, see: David Knight, Insti-
tutions and Social Conflict (Cambridge, Cambridge University Press, 1992). This can
be a problem even in legislatures, where majorities can often enforce institutional
change on minorities, as studies of party government are beginning to emphasize.
See: Gary W. Cox and Matthew D. McCubbins, Legislative Leviathan: Party Govern-
ment in the House (Berkeley, University of California Press, 1994).
76 Cf. Shepsle, op. cit. in note 16.
77 Although some scholars have argued that competition among nation-states or politi-
cal elites tends to select for some kinds of institutions over others, surprisingly little
research has been done on this problem. Cf. Thomas Ertman, Pioneers and Latecomers
(New York, Cambridge University Press, forthcoming); Hilton Root, Fountain of
Privilege (Berkeley, University of California Press, 1994); W. G. Runciman, A Treatise
in Social Theory (Cambridge, Cambridge University Press, 1984); and more generally:
Knight, Institutions and Social Conflict, op. cit. in note 75, ch. 1 and Douglass C. North,
Institutions, Institutional Change and Economic Performance (Cambridge, Cambridge
University Press, 1990).
78 John W. Meyer and Brian Rowan, ‘Institutionalized Organizations: Formal Structure
as Myth and Ceremony’, in John W. Meyer and W. Richard Scott et al., Organisational
Environments: Ritual and Rationality (Newbury Park, Sage, 1992), pp. 21-44; and
George M. Thomas et al., Institutional Structure: Constituting State, Society and the In-
dividual (Beverly Hills, Sage, 1987).
79 There are some notable exceptions, including Fligstein, op. cit. in note 53.
80 For some exceptional works that pay more attention to this dimension of institu-
tionalization, see: Paul J. DiMaggio, ‘Constructing an Organizational Field as a Pro-
fessional Project’ in Powell and DiMaggio, op. cit. in note 49, pp. 267–292; Fligstein,
op. cit in note 53, and Lauren Edelman, ‘Legal Environments and Organizational
Governance,’ American Journal of Sociology 95 (1990), 1401–1440.
81 As Moe and Knight have pointed out, many rational choice analyses are curiously
apolitical. The stress they put on the collective benefits that institutions provide of-
ten seems to mask the degree to which those institutions, like so much else in poli-
tics, emerge out of a struggle for power and resources. Cf. Moe, op. cit in note 72,
and Knight, op. cit in note 73.
82 Cf. Steinmo, op. cit. in note 21, p. 7, and E. E. Schattschneider, The Semi-Sovereign
People (New York, Holt, Rinehart, 1960).
83 Cf. Peter Swenson, ‘Bringing Capital Back In or Social Democracy Reconsidered,’
World Politics 43, 4 (1991), pp. 513–544, and Bo Rothstein, ‘Explaining Swedish Cor-
poratism: The Formative Moment,’ Scandinavian Political Studies 14, 2 (1991), pp. 149–
171.
84 For more on this example, see Hall, op. cit. in note 8, pp. 247–248.
32 MPIFG Discussion Paper 96/6

85 David Kreps, ‘Corporate Culture and Economic Theory,’ in Alt and Shepsle, Perspec-
tives on Positive Political Economy, pp. 90–143.
86 Cf. Geoffrey Garrett and Barry Weingast ‘Ideas, Interests and Institutions: Construct-
ing the European Community’s Internal Market’ in Judith Goldstein and Robert
Keohane (eds.), Ideas and Foreign Policy (Ithaca, Cornell University Press, 1993), pp.
173–206, and Stephen Krasner, op. cit. in note 42.
87 Fritz W. Scharpf, ‘Decision Rules, Decision Styles and Policy Choices,’ Journal of
Theoretical Politics 1, 2 (1989), pp. 149–176. We can read much the same argument
into Robert Putnam’s contention that those regions of Italy with significant past ex-
per-ience of cooperative collective association provided more fertile ground for col-
lective endeavor, even centuries later, than regions without that experience. Cf. Put-
nam, Making Democracy Work: Civic Traditions in Modern Italy (Princeton, Princeton
University Press, 1993).
88 Robert Bates and Barry Weingast, ‘A New Comparative Politics: Integrating Rational
Choice and Interpretivist Perspectives.’ Harvard Center for International Affairs
Working Paper, 1995; Barry Weingast, ‘The Political Foundations of Democracy and
the Rule of Law’ (forthcoming); and John A. Ferejohn, ‘Rationality and Interpreta-
tion: Parliamentary Elections in Early Stuart England’ in Kristen R. Monroe (ed.), The
Economic Approach to Politics (New York, Harper Collins, 1991).
89 For examples of the former, see: Immergut, op. cit. in note 18, and Peter A. Hall,
‘Central Bank Independence and Coordinated Wage Bargaining: The Interdepend-
ence in Germany and Europe,’ German Politics and Society (Autumn 1994); and for
examples of the latter see: Hattam, op. cit. in note 19, and Steinmo, op. cit. in note 21.
90 Cf. Ertman, op. cit. in note 77.
91 Cf. Hattam, op. cit. in note 19.